Shifts in reimbursement trends towards a focus on patient outcomes, total cost of care, and evidence-based medicine are spurring greater reliance on big data to catalyze change. Data are becoming more readily available and originate from a greater variety of sources, including digitized medical records, pharmaceutical records, health information exchanges, and fitness wearables. However, the sensitive nature and multiple uses of these data also raise a number of concerns related to privacy, competition, bias, and intellectual property.

Patient Data, Barriers to Entry, and Market Power

Whether and how access to patient data affects incentives and abilities to foreclose competitors, barriers to entry, or market power is an empirical question. Sharing patient data across provider types (e.g., hospitals, physicians, specialists) or among competitors (e.g., between hospitals) can improve patient care, track disease progression, reduce duplicative testing and costs, help maintain thorough and consistent patient records, and facilitate entry of innovative competitors. Expert analysis can help to determine and quantify the differences between data used as a tool to enhance patient outcomes versus stymieing competition.

Algorithmic Bias

The availability of comprehensive and diverse sources of patient data has encouraged the creation of machine learning algorithms to guide treatment and insurance decisions. Observers have suggested that the quality of the data generated and collected is instrumental to the success of such algorithms, as is the formulation of the algorithm itself. Analyses of alleged bias or disparate outcomes resulting from the use of algorithms benefit from expertise in machine learning, claims data analyses, and the healthcare sector.

Electronic Data, Invasion of Privacy, and Data Breach

The digitization of medical records, as well as the adoption of telemedicine, has created large volumes of electronic data and raised concerns about the susceptibility of health or other personal data to unauthorized disclosure and use. The collection of geolocation data as part of public health initiatives like contact tracing has also contributed to these concerns. Experts in privacy and data breach must evaluate the merits of quantitative methods typically used to calculate damages in these settings, such as contingent valuation surveys and conjoint analysis.

Disease Diagnosis

Machine learning algorithms developed to read medical images can reduce diagnostic errors. Speech recognition software can also check symptoms described by patients against a database of diseases to assist physicians with diagnoses. How should one evaluate concerns regarding the accuracy or precision of such artificial intelligence tools? Can algorithms used for medical diagnoses be patented and are they susceptible to complaints of algorithmic bias? A combination of industry and statistics expertise is necessary to answer these questions.

Software-Driven Medical Devices

There is substantial regulatory uncertainty relating to the treatment of data generated by wearable devices such as smartwatches, activity trackers, and health monitoring devices. These data can help healthcare providers manage patient health more proactively, and ultimately can be welfare enhancing by improving patient outcomes and lowering costs. Questions related to the ownership, protection, and confidentiality of such data can be evaluated using economic approaches tailored to intellectual property issues.

Drug Discovery

Advanced artificial intelligence tools can facilitate and improve several aspects of drug development, including combing through large repositories of biological information to identify promising drug candidates; analyzing data from past clinical trials to find potential new uses for approved drugs; processing data generated by genetic sequencing to create personalized therapies; and investigating viral protein structures.

Expertise in the pharmaceutical industry, and particularly in the analysis of intellectual property issues and of the economic risks of drug development, can help resolve questions regarding data ownership, data protection, and the impact of the FDA’s regulatory decisions.